Dysregulated vascular growth plays a critical role in a variety of inflammatory diseases, in particular psoriasis, delayed type hypersensitivity, contact dermatitis, asthma, multiple sclerosis, restenosis, rheumatoid arthritis and inflammatory bowl disease. Aberrant vascular growth is also involved in neovascular ocular diseases such as age-related macular degeneration and diabetic retinopathy. Additionally, sustained vascular growth is accepted as one hallmark of cancer development (Hanahan, D.; Weinberg, R. A. Cell 2000, 100, 57). While tumours initially grow either as an avascular mass or by co-opting existing host vessels, growth beyond a few mm3 in size is depending on the induction of vessel neogrowth in order to sufficiently provide the tumour with oxygen and nutrients. Induction of angiogenesis is a prerequisite that the tumour surpasses a certain size (the so called angiogenic switch). An intricate signalling interaction network between cancer cells and the tumour microenvironment triggers the induction of vessel growth from existing vasculature. The dependence of tumours on neovascularization has led to a new treatment paradigm in cancer therapy (Ferrara et al. Nature 2005, 438, 967; Carmeliet Nature 2005, 438, 932). Blocking tumour neovascularization by small molecule or antibody-mediated inhibition of relevant signal transduction pathways holds a great promise for extending currently available therapy options.
The development of the cardiovascular system involves two basic stages. In the initial vasculogenesis stage, which only occurs during embryonal development, angioblasts differentiate into endothelial cells which subsequently form a primitive vessel network. The subsequent stage, termed angiogenesis, involves the remodelling of the initial vasculature and sprouting of new vessels (Risau, W. Nature 1997, 386, 671; Jain, R. K. Nat. Med. 2003, 9, 685). Physiologically, angiogenesis occurs in wound healing, muscle growth, the female cycle and in the above mentioned disease states.
It has been found that receptor tyrosine kinases of the vascular endothelial growth factor (VEGF) family and the Tie (tyrosine kinase with immunoglobulin and epidermal growth factor homology domain) receptor tyrosine kinases are essential for both developmental and disease-associated angiogenesis (Ferrara et al Nat. Med. 2003, 9, 669; Dumont et al. Genes Dev. 1994, 8, 1897; Sato et al. Nature 1995, 376, 70).
In adults the Tie2 receptor tyrosine kinase is selectively expressed on endothelial cells (EC) of the adult vasculature (Schlaeger et al. Proc. Nat. Acad. Sci. USA 1997, 94, 3058). Immunohistochemical analysis demonstrated the expression of Tie2 in adult rat tissues undergoing angiogenesis. During ovarian folliculogenesis, Tie2 is expressed in neovessels of the developing corpus luteum. Four endogeneous ligands—angiopoietins 1 to 4—have been identified for the type 1 transmembrane Tie2 (also named Tek) receptor, while no ligands have been identified so far for the Tie1 receptor. Binding of the extracellular Tie2 domain to the C-terminal fibrinogen-like domains of the various angiopoietins leads to significantly different cellular effects. In addition, heterodimerizations between Tie1 and Tie2 receptors have been postulated to influence ligand binding.
Binding of Ang1 to Tie2 expressed on EC induces receptor cross-phosphorylation and kinase activation thus triggering various intracellular signalling pathways. The intracellular C-terminal tail of the Tie2 protein plays a crucial role in Tie2 signalling (Shewchuk et al. Structure 2000, 8, 1105). Upon ligand binding, a conformational change is induced which removes the C-tail out of its inhibitory conformation thus allowing kinase activation by cross-phoshorylation of various Tyr residues in the C-tail, which subsequently function as docking sites for phosphotyrosine-binding (PTB) site possessing down-stream mediators. Cellular effects initiated by Ang1 activation of Tie2 include inhibition of EC apoptosis, stimulation of EC migration and blood vessel reorganization, suppression of inflammatory gene expression and suppression of vascular permeability (Brindle et al. Circ. Res. 2006, 98, 1014). In contrast to VEGF-VEGFR signalling in EC, Ang1 activation of Tie2 does not stimulate EC proliferation in the majority of published assay settings.
The anti-apoptotic effect of Tie2 signalling was shown to be mediated mainly by the PI3K-Akt signalling axis which is activated by binding of the regulatory p85 subunit of PI3K to Y1102 in the Tie2 C-tail (DeBusk et al. Exp. Cell. Res. 2004, 298, 167; Papapetropoulos et al. J. Biol. Chem. 2000, 275, 9102; Kim et al. Circ. Res. 2000, 86, 24). In contrast, the chemotactic response downstream of the activated Tie2 receptor requires crosstalk between PI3K and the adaptor protein Dok-R. Membrane localization of Dok-R via binding of its plekstrin homology (PH) domain to PI3K and simultaneous binding to Y1108 in the Tie2 C-tail via its PTB domain leads to Dok-R phoshorylation and downstream signalling via Nck and Pak-1 (Jones et al. Mol. Cell. Biol. 2003, 23, 2658; Master et al. EMBO J. 2001, 20, 5919). PI3K-mediated recruitment of the adaptor protein ShcA to Y1102 of the Tie2 C-tail is also believed to induce cellular sprouting and motility effects involving activation of endothelial nitric oxide synthase (eNOS), focal adhesion kinase (FAK) and the GTPases RhoA and Rac1. Other downstream mediators of Tie2 signalling include the adaptor protein Grb2, which mediates Erk1/2 stimulation, and the SHP-2 phosphatase.
In conclusion, basal activation of the Tie2 pathway by Ang1 is believed to maintain quiescence and integrity of the endothelium of the adult vasculature by providing a cell survival signal for ECs and by maintaining the integrity of the EC lining of blood vessels (Peters et al. Recent Prog. Horm. Res. 2004, 59, 51).
In contrast to Ang1, Ang2 is not able to activate Tie2 on EC unless Ang2 is present in high concentration or for prolonged periods. However, Ang2 functions as a Tie2 agonist in non-endothelial cells transfected with Tie2. The structural basis for this context-dependence of the Ang2-Tie2 interaction is to date not understood.
In endothelial cells, however, Ang2 functions as Tie2 antagonist and thus blocks the agonistic activity of Ang1 (Maisonpierre et al. Science 1997, 277, 55). Ang2 binding to Tie2 prevents Ang1-mediated Tie2 activation which leads to vessel destabilization and results in vessel regression in the absence of pro-angiogenic stimuli such as VEGF. While Ang1 is widely expressed by periendothelial cells in quiescent vasculature such as pericytes or smooth muscle cells, Ang2 expression occurs in areas of ongoing angiogenesis. Ang2 can be stored in Weibel-Palade bodies in the cytoplasm of EC allowing for a quick vascular response upon stimulation.
Ang1 and Ang2 are expressed in the corpus luteum, with Ang2 localizing to the leading edge of proliferating vessels and Ang1 localizing diffusively behind the leading edge. Ang2 expression is inter alia initiated by hypoxia (Pichiule et al. J. Biol. Chem. 2004, 279, 12171). Ang2 is upregulated in the tumour vasculature and represents one of the earliest tumour markers. In the hypoxic tumour tissue, Ang2 expression induces vessel permeability and—in the presence of e.g. pro-angiogenic VEGF—triggers angiogenesis. After VEGF mediated EC proliferation and vessel sprouting maturation of the newly formed vessels again necessitates Tie2 activation by Ang1. Therefore, a subtle balancing of Tie2 activity plays a pivotal role in the early as well as late stages of neovascularization. These observations render the Tie2 RTK an attractive target for anti-angiogenesis therapy in diseases caused by or associated with dysregulated vascular growth. However, it remains to be shown if targeting the Tie2 pathway alone will be sufficient to achieve efficacious blockade of neovascularization. In certain diseases or disease subtypes it might be necessary or more efficacious to block several angiogenesis-relevant signalling pathways simultaneously.
Various theories have been discussed to explain the differential effects of Ang1 and Ang2 on Tie2 downstream signalling events. Binding of Ang1 and Ang2 in a structurally different manner to the Tie2 ectodomain could induce ligand-specific conformational changes of the intracellular kinase domain explaining different cellular effects. Mutational studies however point toward similar binding sites of Ang1 and Ang2. In contrast, various publications have focussed on different oligomerization states of Ang1 vs. Ang2 as basis for different receptor multimerization states upon ligand binding. Only Ang1 present in its tetramer or higher-order structure initiates Tie2 activation in EC while Ang2 was reported to exist as a homodimer in its native state (Kim et al. J. Biol. Chem. 2005, 280, 20126; Davis et al. Nat. Struc. Biol. 2003, 10, 38; Barton et al. Structure 2005, 13, 825). Finally, specific interactions of Ang1 or Ang2 with additional cell-specific co-receptors could be responsible for the different cellular effects of Ang1 vs. Ang2 binding to Tie2. Interaction of Ang1 with integrin α5β1 has been reported to be essential for certain cellular effects (Carlson et al. J. Biol. Chem. 2001, 276, 26516; Dallabrida et al. Circ. Res. 2005, 96, e8). Integrin α5β1 associates constitutively with Tie2 and increases the receptor's binding affinity for Ang1 resulting in initiation of downstream signalling at lower Ang1 effector concentrations in situations where integrin α5β1 is present. The recently solved crystal structure of the Tie2-Ang2 complex suggests however that neither the oligomerization state nor a different binding mode causes the opposing cellular effects (Barton et al. Nat. Struc. Mol. Biol. 2006, 13, 524).
Ang1-Tie2 signalling plays also a role in the development of the lymphatic system and in lymphatic maintenance and sprouting (Tammela et al. Blood 2005, 105, 4642). An intimate cross-talk between Tie2 and VEGFR-3 signalling in lymphangiogenesis seems to equal the Tie2-KDR cross-talk in blood vessel angiogenesis.
A multitude of studies have underscored the functional significance of Tie2 signalling in the development and maintenance of the vasculature. Disruption of Tie2 function in Tie2−/− transgenic mice leads to early embryonic lethality between days 9.5 and 12.5 as a consequence of vascular abnormalities. Tie2−/− embryos fail to develop the normal vessel hierarchy suggesting a failure of vascular branching and differentiation. The heart and vessels in Tie2−/− embryos show a decreased lining of EC and a loosened interaction between EC and underlying pericyte/smooth muscle cell matrix. Mice lacking functional Ang1 expression and mice overexpressing Ang2 display a phenotype reminiscent of the phenotype of Tie2−/− mice (Suri et al. Cell 1996, 87, 1171). Ang2−/− mice have profound defects in the growth and patterning of lymphatic vasculature and fail to remodel and regress the hyaloid vasculature of the neonatal lens (Gale et al. Dev. Cell 2002, 3, 411). Ang1 rescued the lymphatic defects, but not the vascular remodelling defects. Therefore, Ang2 might function as a Tie2 antagonist in blood vasculature but as a Tie2 agonist in developing lymph vasculature suggesting redundant roles of Ang1 and Ang2 in lymphatic development.
Aberrant activation of the Tie2 pathway is involved in various pathological settings. Activating Tie2 mutations leading to increased ligand-dependent and ligand-independent Tie2 kinase activity cause inherited venous malformations (Vikkula et al. Cell 1996, 87, 1181). Increased Ang1 mRNA and protein levels as well as increased Tie2 activation have been reported in patients with pulmonary hypertension (PH). Increased pulmonary arterial pressure in PH patients results from increased coverage of pulmonary arterioles with smooth muscle cells (Sullivan et al. Proc. Natl. Acad. Sci. USA 2003, 100, 12331). In chronic inflammatory diseases, like in psoriasis, Tie2 and the ligands Ang1 and Ang2 are greatly upregulated in lesions, whereas a significant decrease in expression of Tie2 and ligands occur under anti-psoriatic treatment (Kuroda et al. J. Invest. Dermatol 2001, 116, 713). Direct association of pathogenesis of disease with Tie2 expression has been demonstrated recently in transgenic mice overexpressing Tie2 (Voskas et al. Am. J. Pathol. 2005, 166, 843). In these mice overexpression of Tie2 causes a psoriasis-like phenotype (such as epidermal thickening, rete ridges and lymphocyte infiltration). These skin abnormalities are resolved completely upon suppression of transgene expression, thereby illustrating a complete dependence on Tie2 signalling for disease maintenance and progression. A recent study underscored the connection of the Ang1/Ang2-Tie2 signalling axis to the induction of inflammation (Fiedler et al. Nat. Med. 2006, 12, 235). Inhibition of the Tie2 signalling pathway is therefore expected to be useful in the therapy of a broad range of inflammatory diseases.
Tie2 expression was investigated in human breast cancer specimens and Tie2 expression was found in the vascular endothelium both in normal breast tissue as well as in tumour tissue. The proportion of Tie2-positive microvessels was increased in tumours as compared to normal breast tissue (Peters et al. Br. J. Canc. 1998, 77, 51). However, significant heterogeneity in endothelial Tie2 expression was observed in clinical specimen from a variety of human cancers (Fathers et al. Am. J. Path. 2005, 167, 1753). In contrast, Tie2 and angiopoietins were found to be highly expressed in the cytoplasm of human colorectal adenocarcinoma cells indicating at the potential presence of an autocrine/paracrine growth loop in certain cancers (Nakayama et al. World J. Gastroenterol. 2005, 11, 964). A similar autocrine/paracrine Ang1-Ang2-Tie2 loop was postulated for certain human gastric cancer cell lines (Wang et al. Biochem. Biophys. Res. Comm. 2005, 337, 386).
The relevance of the Ang1-Tie2 signalling axis was challenged with various biochemical techniques. Inhibition of Ang1 expression by an antisense RNA approach resulted in decreased xenograft tumour growth (Shim et al. Int. J. Canc. 2001, 94, 6; Shim et al. Exp. Cell Research 2002, 279, 299). However, other studies report that experimental overexpression of Ang1 in tumour models leads to decreased tumour growth (Hayes et al. Br. J. Canc. 2000, 83, 1154; Hawighorst et al. Am. J. Pathol. 2002, 160, 1381; Stoeltzing et al. Cancer Res. 2003, 63, 3370). The latter results can be rationalized by the ligand's ability to stabilize the endothelial lining of vessels rendering vessels less sensitive for angiogenic stimuli. Interference with the dynamics of Ang1-Tie2 signalling either by over-stimulation or by stimulus deprivation seemingly leads to similar phenotypes.
The pharmacological relevance of inhibiting Tie2 signalling was tested applying various non-small molecule approaches. A peptidic inhibitor of Ang1/2 binding to Tie2 was shown to inhibit Ang1-induced HUVEC migration and angiogenesis induction in an in vivo model (Tournaire et al. EMBO Rep. 2005, 5, 1). Corneal angiogenesis induced by tumour cell conditioned medium was inhibited by a recombinant soluble Tie2 receptor (sTie2) despite the presence of VEGF (Lin et al. J. Clin. Invest. 1997, 100, 2072; see also Singh et al. Biochem. Biophys. Res. Comm. 2005, 332, 194). Gene therapy by adenoviral vector delivered sTie2 was capable of reducing tumour growth rates of a murine mammary carcinoma and a murine melanoma and resulted in reduction of metastasis formation (Lin et al. Proc. Natl. Acad. Sci. USA 1998, 95, 8829). Similar effects were observed with related sTie2 constructs (Siemeister et al. Cancer Res. 1999, 59, 3185) and a Tek-Fc construct (Fathers et al. Am. J. Path. 2005, 167, 1753).
Adenovirus-delivered anti-Tie2 intrabodies were shown to inhibit growth of a human Kaposi's sarcoma and a human colon carcinoma upon peritumoural administration (Popkov et al. Cancer Res. 2005, 65, 972). Histopathological analysis revealed a marked decrease in vessel density in treated vs. control tumours. Phenotypic simultaneous knockout of KDR and Tie2 by an adenovirus delivered intradiabody resulted in significantly higher growth inhibition of a human melanoma xenograft model than KDR knockout alone (Jendreyko et al. Proc. Natl. Acad. Sci. USA 2005, 102, 8293). Similarly, the bispecific Tie2-KDR intradiabody was more active in an in vitro EC tube formation inhibition assay than the two monospecific intrabodies alone (Jendreyko et al. J. Biol. Chem. 2003, 278, 47812). Systematic treatment of tumour-bearing mice with Ang2-blocking antibodies and peptide-Fc fusion proteins led to tumour stasis and elimination of tumour burden in a subset of animals (Oliner et al. Cancer Cell 2004, 6, 507). For a recent report on an immunization approach, see Luo et al. Clin. Cancer Res. 2006, 12, 1813.
However, from the above studies using biochemical techniques to interfere with Tie2 signalling it is not clear, whether similar phenotypes will be observed with small molecule inhibitors of the Tie2 kinase activity. Small molecule inhibitors of kinases by definition block only those cellular effects which are mediated by the receptor's kinase activity and not those which might involve the kinase only as a co-receptor or scaffolding component in multi-enzyme complexes. So far, studies describing in vivo pharmacodynamic effects of small molecule Tie2 inhibitors are rare (Scharpfenecker et al. J. Cell Sci. 2005, 118, 771; J. M. Chen, Medicinal Chemistry and High Speed Synthesis—The Tie-2 story; presentation held at the Centennial AACR, April 2007, Los Angeles, U.S.A.) It remains to be shown that small molecule inhibitors of the Tie2 kinase will be as efficacious in inhibiting angiogenesis as e.g. ligand antibodies, soluble decoy receptors or receptor intrabodies. As discussed above, in certain settings inhibition of Tie2 signalling alone might not be sufficient to induce an adequate antiangiogenic effect. Simultaneous inhibition of several angiogenesis relevant signalling pathways could overcome such inadequacies. In conclusion, there is a great need for novel chemotypes for small molecule inhibitors of the Tie2 kinase. Fine tuning of additive anti-angiogenic activities as well as pharmacokinetic parameters such as e.g. solubility, membrane permeability, tissue distribution and metabolism will finally allow for chosing compounds of accurate profiles for various diseases caused by or associated with dysregulated vascular growth.